Study On FECG Acquisition System Based On VMD Method

FECG(Fetal Electrocardiogram)is a kind of electrical signal that records the activities of the fetal heart.It can reflect the health status of the foetus and provide protection for foetal development.The indirect measurement method of FECG isolates the FECG from the AECG(Abdominal Electrocardiogram)and is non-invasive and easy to operate.However,indirect measurements are susceptible to EMG noise and power frequency noise due to the limited of signal propagation path,so it is difficult to directly observe the effective signal when it reaches the acquisition end.FECG signal amplitude is weak,collection difficulty is high,and the abdominal wall signal components are complex,so FECG filtering extraction algorithms have been the focus of research.This thesis focuses on FECG acquisition and filtering.The goal is to study and design a FECG acquisition system capable of collecting a well-defined waveform of the raw AECG,based on which a filtering algorithm can be further investigated to extract the desired FECG.The contents of this thesis are as follows.1.ECG denoising algorithm based on VMD(variational mode decomposition)method and smooth filtering is proposed.In order to solve the problem of difficult trade-offs of IMFs(Intrinsic Mode Functions)in the current the VMD for the ECG noise reduction an improved VMD method is proposed.The noisy ECG is first decomposed into different IMFs using the VMD method.The noise-dominated and signal-dominated IMFs are determined by the central frequency of each modal component and the cross correlation between the IMF and the raw ECG signal.The number of sampling points corresponding to the heartbeat frequency is extracted using the IMF of the central frequency in the range of the medical heartbeat frequency.Depending on the heartbeat frequency,noise-dominated and signal-dominated IMFs are smoothed separately.Finally,the processed IMFs are used to reconstruct the ECG to complete the baseline drift and EMG noise removal.Experimental results show that the proposed method performs well in denoising.2.A fast baseline drift extraction algorithm based on the mean filter is proposed.In this thesis,the unavoidable baseline drift noise in ECG noise is removed by performing R-wave calibration with a low-computation mean filter,and then the reference point is chosen for baseline fitting to achieve the removal of baseline drift.The proposed method is highly resistant to power-frequency interference and myoelectric interference,and can be deployed on microcontrollers for ECG acquisition,enabling real-time denoising with minimal computational effort.3.FECG acquisition system based on instrumental amplifiers and Qt framework is developed.First,the hardware design of the FECG acquisition device is performed in accordance with the key requirements and standards of ECG machines.The UI software is designed in the framework of Qt.With a sampling rate of 2KHz and a bandwidth of 0 to150 Hz,the noise of the system can be kept below 2u V,the FECG signal can be collected above 10 u V,and the sampling rate can be dynamically adjusted.The system uses three transmission modes USB,Ethernet and WIFI to meet various application scenarios.Using a precision signal source,the gain error of the system is less than 0.1%,which can meet the needs of fetal ECG acquisition.